Multiband noisy spectrum sensing with codebooks
Nancy Ronquillo, Tara Javidi · 2016
Cognitive radio technology aims to meet the demand for higher data rates in wireless communications by allowing secondary users to opportunistically locate and utilize otherwise under-utilized spectrum resources [1]. This paper focuses on the cases when a secondary user monitors the occupancy of frequency subbands by using a multiband spectrum sensing approach in which multiple frequency subbands are simultaneously searched in groups. This paper establishes a direct connection between multiband spectrum sensing and the problem of encoding over a Multiple Access Channel (MAC) with feedback. As a main consequence, this approach allows for the characterization of fundamental limits on the speed and the reliability of optimal multiband spectrum sensing. In particular, this paper considers a selection of specialized non-adaptive, and adaptive codes with previously guaranteed performance that can be used as sensing schemes. In the special case where there is exactly one vacant frequency subband, a variant of the posterior matching coding scheme [2] for point-to-point channels is proposed. Known results on the performance of this coding scheme are then used to upper bound the expected minimum number of measurements needed to locate the vacant subband.